Information processing method and device

Abstract

Embodiments of the invention provide an information processing method and apparatus. The information processing method comprises the steps of determining initial insurance information in response to an insurance analysis request submitted for a target insurance item; the initial insurance information is input into a large language model, and the large language model comprises an information updating unit and a multi-path recall unit; updating the initial insurance information into target insurance information by using the information updating unit, and recalling multiple pieces of insurance sub-information associated with the insurance analysis request in the target insurance information through the multi-path recall unit; and inputting the multiple pieces of insurance sub-information and the initial insurance information into a target inference model for processing to obtain target insurance sub-information associated with the insurance analysis request.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of artificial intelligence technology, and more particularly to information processing methods and devices. Background Art

[0002] With the development of computer and internet technologies, information processing efficiency has increased significantly. However, in insurance project scenarios, due to the wide range of information coverage and lack of unified standards, manual intervention is required to normalize liability factors. Manual processing is not only time-consuming and costly, but also difficult to cope with the changing and large external data environment. Existing technologies have strategies for parsing insurance product-related information through large models and NER (Named Entity Recognition). While these strategies can address the high cost and low efficiency of manual processing, a single large-model parsing approach cannot handle the diverse range of insurance clause data, and NER methods lack a deep understanding of complex semantics. Consequently, both methods suffer from insufficient data accuracy and recall. Therefore, an effective solution is urgently needed to address these issues. Summary of the Invention

[0003] In view of this, embodiments of this specification provide an information processing method. One or more embodiments of this specification also relate to an information processing apparatus, a computing device, a computer-readable storage medium, and a computer program product to address technical deficiencies in the prior art.

[0004] According to a first aspect of the embodiments of this specification, there is provided an information processing method, including: determining initial insurance information in response to an insurance analysis request submitted for a target insurance program; Inputting the initial insurance information into a large language model, wherein the large language model includes an information updating unit and a multi-way recall unit; Using the information updating unit to update the initial insurance information to target insurance information, and using the multi-way recalling unit to recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information; The multiple insurance sub-information and the initial insurance information are input into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request.

[0005] Optionally, determining initial insurance information in response to an insurance analysis request submitted for a target insurance program includes: receiving an insurance analysis request submitted for a target insurance project, and obtaining insurance information associated with the target insurance project according to the insurance analysis request; The insurance information is preprocessed and formatted to obtain initial insurance information.

[0006] Optionally, the updating the initial insurance information to target insurance information by using the information updating unit includes: inputting the initial insurance information into the information updating unit; The information updating unit is used to perform semantic enhancement and information rewriting on the initial insurance information to obtain target insurance information.

[0007] Optionally, recalling the plurality of insurance sub-information associated with the insurance analysis request from the target insurance information by the multi-channel recall unit includes: Performing information retrieval on the target insurance information using a recall strategy preset by each of the multiple recall units to obtain at least one initial insurance sub-information corresponding to each recall unit; Fusing at least one initial insurance sub-information corresponding to each recall unit to obtain intermediate insurance sub-information corresponding to each recall unit; A plurality of insurance sub-information is formed based on the intermediate insurance sub-information corresponding to each recall unit, wherein the recall strategy preset for each recall unit is different.

[0008] Optionally, inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request includes: calculating a correlation degree between each insurance sub-information in the plurality of insurance sub-information and the insurance analysis request; Sort the multiple insurance sub-information according to the relevance to obtain a sub-information list; The sub-information list and the initial insurance information are input into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request.

[0009] Optionally, after the step of inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request is performed, the method further includes: Inputting the initial insurance information and the target insurance sub-information into an evaluation model for processing to obtain an information evaluation result; If the information evaluation result is passed, feeding back the target insurance sub-information to the client that submitted the insurance analysis request; If the information evaluation result is failed, supplementary information is determined based on the information evaluation result, and the supplementary information is merged with the target insurance sub-information to obtain global insurance sub-information, which is then fed back to the client that submitted the insurance analysis request.

[0010] Optionally, after the step of inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request is performed, the method further includes: receiving an adjustment request for the target insurance sub-information feedback, and constructing model optimization information according to the adjustment request; The large language model and the target reasoning model are optimized based on the model optimization information until a large language model and a target reasoning model that meet optimization conditions are obtained.

[0011] According to a second aspect of the embodiments of this specification, another information processing method is provided, including: Receiving an insurance analysis request submitted by a user for a target insurance project, and obtaining initial insurance information associated with the target insurance project according to the insurance analysis request; Inputting the initial insurance information into a large language model, wherein the large language model includes an information updating unit and a multi-way recall unit; Using the information updating unit to update the initial insurance information to target insurance information, and using the multi-way recalling unit to recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information; Inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request; Construct explanatory information corresponding to the target insurance sub-information, and display the target insurance sub-information and the explanatory information to the user.

[0012] According to a third aspect of the embodiments of this specification, there is provided an information processing device, including: a determination module configured to determine initial insurance information in response to an insurance analysis request submitted for a target insurance program; an input module configured to input the initial insurance information into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit; an updating module configured to update the initial insurance information to target insurance information using the information updating unit, and recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information using the multi-way recall unit; The processing module is configured to input the multiple insurance sub-information and the initial insurance information into a target reasoning model for processing, and obtain target insurance sub-information associated with the insurance analysis request.

[0013] According to a fourth aspect of the embodiments of this specification, another information processing device is provided, including: a request receiving module configured to receive an insurance analysis request submitted by a user for a target insurance project, and obtain initial insurance information associated with the target insurance project according to the insurance analysis request; An input model module is configured to input the initial insurance information into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit; an information updating module configured to update the initial insurance information to target insurance information using the information updating unit, and recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information using the multi-way recall unit; an information processing module configured to input the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing, and obtain target insurance sub-information associated with the insurance analysis request; The information construction module is configured to construct explanation information corresponding to the target insurance sub-information and display the target insurance sub-information and the explanation information to the user.

[0014] According to a fifth aspect of the embodiments of this specification, there is provided a computing device, including: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the above-mentioned information processing method are implemented.

[0015] According to a sixth aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned information processing method are implemented.

[0016] According to a seventh aspect of the embodiments of this specification, a computer program product is provided, comprising a computer program or instructions, which implement the steps of the above-mentioned information processing method when executed by a processor.

[0017] The information processing method provided in this embodiment can determine initial insurance information in response to an insurance analysis request submitted for a target insurance project in order to adapt to more insurance products and save human resources while improving the efficiency of insurance information analysis. At this time, the initial insurance information can be input into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit. In order to ensure the comprehensiveness of the analysis results, the initial insurance information can be updated to the target insurance information using the information update unit to avoid interference caused by redundant information. After that, the multiple insurance sub-information associated with the insurance analysis request can be recalled in the target insurance information through the multi-way recall unit, thereby ensuring the comprehensiveness of the recall results. Thereafter, the multiple insurance sub-information and the initial insurance information can be input into the target reasoning model for processing, so that the reasoning model can be further analyzed at the semantic level to obtain the target insurance sub-information associated with the insurance analysis request. When analyzing insurance information, the large language model and the reasoning model can be combined to complete the information analysis, which not only ensures the accuracy of the analysis but also avoids the confusion caused by redundant information, thereby facilitating the use of downstream businesses. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of an information processing method provided by one embodiment of this specification; Figure 2 is a flowchart of another information processing method provided by one embodiment of this specification; Figure 3 This is a flowchart of a processing process of an information processing method provided by one embodiment of this specification; Figure 4 This is a schematic diagram of the structure of an information processing device provided by one embodiment of this specification; Figure 5 is a structural diagram of another information processing device provided by an embodiment of this specification; Figure 6 This is a structural block diagram of a computing device provided by one embodiment of this specification. DETAILED DESCRIPTION

[0019] The following description sets forth many specific details to facilitate a thorough understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the scope of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0020] The terms used in one or more embodiments of this specification are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of this specification. The singular forms "a," "the," and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0021] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0022] In addition, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0023] First, the terms involved in one or more embodiments of this specification are explained.

[0024] Large Language Model (LLM) is an AI model based on deep learning. It uses unsupervised training on massive amounts of text data (such as books, articles, and code) to master the ability to understand and generate natural language. Its core is to simulate human language patterns and support a variety of tasks, including text generation, question answering, translation, summarization, and logical reasoning.

[0025] RAG (Retrieval-Augmented Generation) is a technical framework that combines information retrieval and generative models, aiming to address the shortcomings of large language models (LLMs). It retrieves relevant information from external knowledge bases (such as databases, documents, and web pages) and inputs it into the LLM as context, thereby improving the accuracy, real-time nature, and interpretability of generated results.

[0026] In this specification, an information processing method is provided. This specification also relates to an information processing apparatus, a computing device, a computer-readable storage medium, and a computer program product, which are described in detail one by one in the following embodiments.

[0027] See also Figure 1 , Figure 1 A flowchart of an information processing method provided according to an embodiment of this specification is shown, which specifically includes the following steps.

[0028] Step S102 : determining initial insurance information in response to an insurance analysis request submitted for a target insurance program.

[0029] The information processing method provided in this embodiment can be applied to any insurance analysis scenario, such as auto insurance, disease insurance, medical insurance, and pension insurance. This embodiment does not impose any limitations. The method analyzes insurance information related to an insurance item to determine target insurance sub-information, namely, the liability factors associated with the insurance item. This helps users understand the insurance content and facilitates subsequent use. Specifically, the liability factors refer to information within the insurance item that users need to pay attention to, such as the types of diseases covered in disease insurance.

[0030] Specifically, the insurance analysis request specifically refers to the request submitted by the insurance analysis demander, and the initial insurance information specifically refers to the insurance information of the associated target insurance project, such as the insurance-related name, liability name, liability description, insurance scope and other information. During specific implementation, the information content contained therein can be set according to actual needs, and this embodiment does not make any restrictions here.

[0031] The information processing method provided in this embodiment can determine initial insurance information in response to an insurance analysis request submitted for a target insurance project in order to adapt to more insurance products and save human resources while improving the efficiency of insurance information analysis. At this time, the initial insurance information can be input into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit. In order to ensure the comprehensiveness of the analysis results, the initial insurance information can be updated to the target insurance information using the information update unit to avoid interference caused by redundant information. After that, the multiple insurance sub-information associated with the insurance analysis request can be recalled from the target insurance information through the multi-way recall unit, thereby ensuring the comprehensiveness of the recall results. Thereafter, the multiple insurance sub-information and the initial insurance information can be input into the target reasoning model for processing, so that the reasoning model can be further analyzed at the semantic level to obtain the target insurance sub-information associated with the insurance analysis request. When analyzing insurance information, the large language model and the reasoning model can be combined to complete the information analysis, which not only ensures the accuracy of the analysis but also avoids the confusion caused by redundant information, thereby facilitating the use of downstream businesses.

[0032] Furthermore, in order to facilitate subsequent processing, the insurance information may be pre-processed and format converted. In this embodiment, the specific implementation is as follows: An insurance analysis request submitted for a target insurance project is received, and insurance information associated with the target insurance project is obtained according to the insurance analysis request; the insurance information is preprocessed, and the format of the preprocessed insurance information is converted to obtain initial insurance information.

[0033] Specifically, insurance information specifically refers to obtaining original insurance information that has not been processed, and format conversion specifically refers to converting the insurance information into a unified format for easy subsequent processing and use.

[0034] Based on this, before determining the initial insurance information, an insurance analysis request submitted for the target insurance project can be received, and the insurance information associated with the target insurance project can be obtained according to the insurance analysis request; at this time, in order to improve the information quality, the insurance information can be preprocessed, and the format of the preprocessed insurance information can be converted to obtain the initial insurance information for subsequent use.

[0035] For example, when it is necessary to analyze the insurance item "critical illness insurance", in order to determine the types of diseases covered by the insurance item, you can first obtain the insurance information of the "critical illness insurance", such as the insurance name, insurance instructions, terms, liability description and other multi-dimensional information. At this time, the obtained insurance information can be cleaned and deduplicated to delete redundant and repeated information. After pre-processing the insurance information, its format can also be converted to determine the initial insurance information based on the conversion results, so as to facilitate the subsequent processing of high-quality initial insurance information and accurately extract information about the disease type from the insurance information.

[0036] In summary, by automatically acquiring and preprocessing insurance information, combined with standardized format conversion, data consistency and processing efficiency are effectively improved, providing high-quality data support for subsequent intelligent decision-making such as risk assessment and personalized recommendations, significantly optimizing insurance business processes, reducing manual intervention costs, enhancing service accuracy and customer experience, and helping the insurance industry achieve digital transformation and enhance core competitiveness.

[0037] Step S104: inputting the initial insurance information into a large language model, wherein the large language model includes an information updating unit and a multi-way recall unit.

[0038] Step S106: Utilize the information updating unit to update the initial insurance information into target insurance information, and utilize the multi-way recalling unit to recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information.

[0039] Specifically, after obtaining the initial insurance information corresponding to the target insurance item, the initial insurance information can be input into a large language model comprising an information update unit and a multi-way recall unit. The information update unit is used to update the initial insurance information, such as performing semantic enhancement and rewriting, to ensure higher-quality and more accurate output insurance information. The multi-way recall unit is used to recall insurance sub-information related to the insurance analysis request from the insurance information using a variety of different recall strategies, thereby facilitating subsequent analysis and use.

[0040] Therefore, after the initial insurance information is input into the large language model, the information update unit can be used to update the initial insurance information to the target insurance information. On this basis, the multiple insurance sub-information associated with the insurance analysis request can be recalled in the target insurance information through the multi-way recall unit, so as to subsequently determine the target insurance sub-information that matches the insurance analysis request from the multiple insurance sub-information.

[0041] The large language model specifically refers to the large model used to process and recall information. The information update unit specifically refers to the unit that rewrites and semantically enhances the initial insurance information. The multi-way recall unit specifically refers to a network composed of multiple sub-recall units, each of which uses a different recall strategy to recall insurance sub-information within the insurance information for subsequent analysis. The target insurance information is the insurance information obtained after updating the initial insurance information, and the insurance sub-information is the insurance sub-information obtained from the target insurance information after recall that matches the insurance analysis request.

[0042] Furthermore, the insurance information can be semantically enhanced and rewritten through the information updating unit. In this embodiment, the specific implementation is as follows: The initial insurance information is input into the information updating unit; and the initial insurance information is semantically enhanced and rewritten using the information updating unit to obtain target insurance information.

[0043] Based on this, after the initial insurance information is input into the large language model, the initial insurance information can be input into the information update unit; the information update unit is used to perform semantic enhancement and information rewriting on the initial insurance information to improve the accuracy and depth of information expression, and thus obtain the target insurance information.

[0044] In summary, the initial insurance information is semantically enhanced and rewritten through the information update unit, which effectively improves the semantic accuracy and naturalness of data expression, solves the problems of information fragmentation and vague expression in traditional methods, enhances risk identification and clause matching capabilities, reduces manual verification costs, and provides dynamically adaptive and logically rigorous target insurance information for insurance business, significantly optimizing intelligent service efficiency and customer satisfaction.

[0045] Furthermore, the multi-channel recall unit can be completed according to different strategies when recalling insurance sub-information. In this embodiment, the specific implementation is as follows: The target insurance information is retrieved through the recall strategy preset by each recall unit in the multiple recall units to obtain at least one initial insurance sub-information corresponding to each recall unit; the at least one initial insurance sub-information corresponding to each recall unit is fused to obtain the intermediate insurance sub-information corresponding to each recall unit; and multiple insurance sub-information are composed based on the intermediate insurance sub-information corresponding to each recall unit, wherein the recall strategy preset by each recall unit is different.

[0046] Specifically, the preset recall strategy specifically refers to the strategy of recall information pre-set by each recall unit. Different recall units correspond to different recall strategies, such as BM25 recall, vector recall, collaborative filtering recall, matrix decomposition recall, etc. During specific implementation, it can be set according to actual needs. This embodiment does not make any restrictions here. The multi-channel recall unit in this embodiment includes at least two-channel recall processing. Correspondingly, the intermediate insurance sub-information specifically refers to the result obtained after fusing the insurance sub-information recalled by each recall unit.

[0047] Based on this, when performing multi-channel recall, the target insurance information can be retrieved through the recall strategy preset by each recall unit in the multi-channel recall unit to obtain at least one initial insurance sub-information corresponding to each recall unit; thereafter, the at least one initial insurance sub-information corresponding to each recall unit can be fused to obtain the intermediate insurance sub-information corresponding to each recall unit; and then, multiple insurance sub-information can be composed based on the intermediate insurance sub-information corresponding to each recall unit, wherein the recall strategy preset by each recall unit is different, thereby realizing information recall through different recall strategies, maximizing the coverage of relevant information, and utilizing the fusion strategy to improve the comprehensiveness of the information.

[0048] Continuing with the previous example, after obtaining the initial insurance information corresponding to "critical illness insurance," we can input this information into the large language model for semantic enhancement and rewriting. For example, if "critical illness insurance" lacks liability description information, the large language model can be used to supplement the liability description of "critical illness insurance" to improve the expressiveness and accuracy of the information. The supplemented liability description information is {Critical illness insurance refers to the coverage amount for the insured's major illness during the policy period...}. This supplemented target insurance information can then be recalled through multiple recall units, each employing a different recall strategy. The initial insurance sub-information recalled by each recall unit is then obtained.

[0049] For example, if the multi-channel recall unit includes three-channel recall units, then the insurance sub-information B11, B12, and B13 recalled from the target insurance information by the first recall unit can be obtained according to the insurance analysis request {disease types covered by the insurance item}. Similarly, the insurance sub-information B21, B22, and B23 recalled from the target insurance information by the second recall unit can be obtained, and the insurance sub-information B31, B32, and B33 recalled from the target insurance information by the third recall unit can be obtained. Furthermore, by fusing the insurance sub-information B11, B12, and B13, the fused insurance sub-information B1 corresponding to the first recall unit, the fused insurance sub-information B2 corresponding to the second recall unit, and the fused insurance sub-information B3 corresponding to the third recall unit can be obtained. This can then be used as a basis to determine the disease types covered by the insurance item for downstream use.

[0050] In summary, through the collaborative retrieval and intelligent fusion processing of differentiated strategies of multiple recall units, multi-dimensional coverage and precise association of target insurance information can be achieved, effectively improving recall efficiency and information relevance, reducing redundant data interference, forming a diversified insurance sub-information set, and significantly enhancing risk matching and personalized recommendation capabilities. It provides dynamically adaptive and highly robust intelligent decision-making support for complex business scenarios, and comprehensively improves the precision of insurance services and operational efficiency.

[0051] Step S108: Input the multiple insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request.

[0052] Specifically, after recalling the multiple insurance sub-information corresponding to the associated insurance analysis request through the multi-channel recall unit, in order to ensure the accuracy of the information, the multiple insurance sub-information and the initial insurance information can be input into the target reasoning model for processing, so that the target reasoning model can perform sub-information analysis from a semantic level, thereby obtaining the target insurance sub-information associated with the insurance analysis request.

[0053] Among them, the target reasoning model specifically refers to a model that matches multiple insurance sub-information from a semantic level and selects target insurance sub-information from them. The target insurance sub-information specifically refers to the sub-information that matches the insurance analysis request. For example, the recalled multiple insurance sub-information includes the names of lung cancer, rectal cancer, gastric cancer, laryngeal cancer, and breast cancer diseases, and the insurance analysis request needs to confirm whether malignant lung tumors are within the scope of insurance reimbursement. At this time, semantic matching can be used to feedback relevant insurance descriptions about lung cancer, thereby achieving semantic matching of malignant lung tumors with lung cancer, and accurately feedback to users or for downstream business.

[0054] Furthermore, the insurance sub-information can be sorted before being processed by the model. In this embodiment, the specific implementation is as follows: Calculate the degree of association between each insurance sub-information in the multiple insurance sub-information and the insurance analysis request; sort the multiple insurance sub-information according to the degree of association to obtain a sub-information list; input the sub-information list and the initial insurance information into the target reasoning model for processing to obtain the target insurance sub-information associated with the insurance analysis request.

[0055] Specifically, the correlation refers to a value that indicates the strength of the association between the insurance sub-information and the insurance analysis request, with a higher correlation indicating a stronger relationship. The sub-information list refers to a queue obtained by sorting multiple insurance sub-information items by correlation, where the information can be sorted from highest to lowest correlation.

[0056] Based on this, before performing reasoning, the correlation between each insurance sub-information in multiple insurance sub-information and the insurance analysis request is calculated; the multiple insurance sub-information is sorted according to the correlation to obtain a sub-information list; thereafter, the sub-information list and the initial insurance information can be input into the target reasoning model for processing, and then the target insurance sub-information associated with the insurance analysis request can be obtained, and precise sorting can be achieved through recall results, which can ensure the accuracy of information sorting so that subsequent semantic analysis can be more accurate.

[0057] Continuing with the above example, after obtaining the fused insurance sub-information B1 corresponding to the first recall unit, the fused insurance sub-information B2 corresponding to the second recall unit, and the fused insurance sub-information B3 corresponding to the third recall unit, the fused insurance sub-information can be sorted according to the correlation between each fused insurance sub-information and the insurance analysis request. According to the sorting result, it is determined that B1>B3>B2. Then, the sorted fused insurance sub-information and the initial insurance information can be input into the target reasoning model for semantic matching to analyze the complex relationship in the insurance terms and infer a more context-relevant liability factor. According to the reasoning results, it is determined that the types of diseases covered by the insurance items in the "critical illness insurance" include "malignant tumors, acute myocardial infarction, sequelae of cerebral stroke, major organ transplant surgery, coronary artery bypass surgery, and end-stage renal disease", so that downstream businesses can complete further analysis and processing of the insurance according to this information.

[0058] In summary, through quantitative evaluation of association metrics and dynamic sorting mechanism, accurate screening of insurance sub-information can be achieved, and combined with the target reasoning model, the semantic relationship of data can be deeply mined, which effectively improves the accuracy of information matching and response efficiency, breaks through the limitations of traditional static retrieval mode, reduces the need for manual intervention, and provides intelligent and explainable decision-making basis for complex insurance scenarios. It significantly enhances risk identification capabilities and service customization levels, and drives the intelligent upgrade of insurance business.

[0059] Furthermore, after the model outputs the target insurance sub-information, it can also be calibrated through the evaluation model. In this embodiment, the specific implementation method is as follows: The initial insurance information and the target insurance sub-information are input into the evaluation model for processing to obtain an information evaluation result; if the information evaluation result is passed, the target insurance sub-information is fed back to the client that submitted the insurance analysis request; if the information evaluation result is failed, supplementary information is determined based on the information evaluation result, and the supplementary information and the target insurance sub-information are merged to obtain global insurance sub-information, and the global insurance sub-information is fed back to the client that submitted the insurance analysis request.

[0060] Specifically, the evaluation model refers to a discriminant model for determining whether the target insurance sub-information meets the intent of the insurance analysis request, and the supplementary information refers to the information of the related target insurance items supplemented to the target insurance sub-information when the target insurance sub-information is incomplete and does not meet downstream usage requirements.

[0061] Based on this, after obtaining the target insurance sub-information, in order to avoid the inaccuracy of the target insurance sub-information, the initial insurance information and the target insurance sub-information can be input into the evaluation model for processing to obtain the information evaluation result; if the information evaluation result is passed, the target insurance sub-information can be directly fed back to the client that submitted the insurance analysis request; if the information evaluation result is failed, supplementary information can be determined based on the information evaluation result, and the supplementary information and the target insurance sub-information can be integrated to obtain the global insurance sub-information, and then the global insurance sub-information can be fed back to the client that submitted the insurance analysis request.

[0062] Continuing with the above example, when the target insurance sub-information obtained is "malignant tumors, acute myocardial infarction, sequelae of cerebral stroke, major organ transplant surgery, coronary artery bypass surgery, and end-stage renal disease", it is input into the evaluation model for processing, and it is determined that other diseases besides the above diseases are recorded in the insurance information. At this time, the disease information to be supplemented can be extracted, such as "leukemia, lymphoma". After that, the supplemented information can be integrated with the target insurance sub-information obtained above, thereby facilitating the use of accurate insurance sub-information downstream.

[0063] In summary, by building a dynamic closed-loop mechanism through the evaluation model, multi-dimensional verification and intelligent completion of insurance information can be achieved, which can improve data quality while reducing the cost of manual intervention; based on the global optimization strategy, the supplementary information is integrated to significantly enhance the integrity of the analysis results and the reliability of decision-making, forming an end-to-end automated risk assessment system, effectively supporting the precise service output in complex scenarios, and driving the intelligent upgrade of insurance business and the leap in operational efficiency.

[0064] In addition, the feedback result of the target insurance sub-information can optimize the model. In this embodiment, the specific implementation method is as follows: An adjustment request for the target insurance sub-information feedback is received, and model optimization information is constructed according to the adjustment request; and the large language model and the target reasoning model are optimized based on the model optimization information until a large language model and a target reasoning model that meet the optimization conditions are obtained.

[0065] Specifically, the model optimization information refers to the information for optimizing the large language model and the target inference model, and the optimization condition refers to the condition for stopping the optimization of the model, including but not limited to the loss value comparison condition, the number of iterations condition or the verification set verification condition. This embodiment does not make any restrictions here.

[0066] Based on this, in order to make the model have more accurate prediction accuracy, it can receive adjustment requests for target insurance sub-information feedback. At this time, model optimization information can be constructed according to the adjustment request. On this basis, the large language model and the target reasoning model can be optimized based on the model optimization information until the large language model and target reasoning model that meet the optimization conditions are obtained, and then deployed to the business scenario for use.

[0067] In summary, by receiving feedback adjustment requests and building model optimization information, dynamic iterative optimization of the large language model and the target inference model can be achieved, forming a multimodal feedback-driven data closed loop, effectively improving the model's generalization ability and response accuracy, strengthening the system's adaptability and stability, and significantly reducing the cost of manual parameter adjustment. This allows for a leap in end-to-end intelligent decision-making capabilities in complex insurance scenarios, driving the intelligent upgrade and value transformation of industry model services.

[0068] The information processing method provided in this embodiment can determine initial insurance information in response to an insurance analysis request submitted for a target insurance project in order to adapt to more insurance products and save human resources while improving the efficiency of insurance information analysis. At this time, the initial insurance information can be input into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit. In order to ensure the comprehensiveness of the analysis results, the initial insurance information can be updated to the target insurance information using the information update unit to avoid interference caused by redundant information. After that, the multiple insurance sub-information associated with the insurance analysis request can be recalled from the target insurance information through the multi-way recall unit, thereby ensuring the comprehensiveness of the recall results. Thereafter, the multiple insurance sub-information and the initial insurance information can be input into the target reasoning model for processing, so that the reasoning model can be further analyzed at the semantic level to obtain the target insurance sub-information associated with the insurance analysis request. When analyzing insurance information, the large language model and the reasoning model can be combined to complete the information analysis, which not only ensures the accuracy of the analysis but also avoids the confusion caused by redundant information, thereby facilitating the use of downstream businesses.

[0069] See also Figure 2 , Figure 2A flowchart of another information processing method provided according to an embodiment of this specification is shown, which specifically includes the following steps.

[0070] Step S202: receiving an insurance analysis request submitted by a user for a target insurance project, and obtaining initial insurance information associated with the target insurance project according to the insurance analysis request.

[0071] Step S204: input the initial insurance information into a large language model, wherein the large language model includes an information updating unit and a multi-way recall unit.

[0072] Step S206: Utilize the information updating unit to update the initial insurance information into target insurance information, and utilize the multi-way recalling unit to recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information.

[0073] Step S208: Input the multiple insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request.

[0074] Step S210: constructing explanation information corresponding to the target insurance sub-information, and displaying the target insurance sub-information and the explanation information to the user.

[0075] This embodiment provides another information processing method, wherein any content not described in detail may refer to the same or corresponding description in the above embodiment, and this embodiment does not impose any limitation thereto. The explanation information specifically refers to information that explains the target insurance sub-information using a large language model.

[0076] The information processing method provided in this embodiment can determine initial insurance information in response to an insurance analysis request submitted for a target insurance project in order to adapt to more insurance products and save human resources while improving the efficiency of insurance information analysis. At this time, the initial insurance information can be input into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit. In order to ensure the comprehensiveness of the analysis results, the initial insurance information can be updated to the target insurance information using the information update unit to avoid interference caused by redundant information. After that, the multiple insurance sub-information associated with the insurance analysis request can be recalled from the target insurance information through the multi-way recall unit, thereby ensuring the comprehensiveness of the recall results. Thereafter, the multiple insurance sub-information and the initial insurance information can be input into the target reasoning model for processing, so that the reasoning model can be further analyzed at the semantic level to obtain the target insurance sub-information associated with the insurance analysis request. When analyzing insurance information, the large language model and the reasoning model can be combined to complete the information analysis, which not only ensures the accuracy of the analysis but also avoids the confusion caused by redundant information, thereby facilitating the use of downstream businesses.

[0077] The following combined Figure 3 , taking the application of the information processing method provided in this specification in the disease insurance information processing scenario as an example, the information processing method is further explained. Figure 3 A flowchart of a processing process of an information processing method provided by an embodiment of this specification is shown, which specifically includes the following steps.

[0078] Step S302: receiving an insurance analysis request submitted for a target insurance project, and obtaining insurance information associated with the target insurance project according to the insurance analysis request.

[0079] Step S304: pre-process the insurance information and convert the format of the pre-processed insurance information to obtain initial insurance information.

[0080] Step S306: input the initial insurance information into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit.

[0081] Step S308: Using the information updating unit, semantic enhancement and information rewriting are performed on the initial insurance information to obtain target insurance information.

[0082] Step S310 , performing information retrieval on target insurance information through the recall strategy preset in each of the multiple recall units, and obtaining at least one initial insurance sub-information corresponding to each recall unit.

[0083] Step S312: fuse at least one initial insurance sub-information corresponding to each recall unit to obtain intermediate insurance sub-information corresponding to each recall unit.

[0084] Step S314 , composing a plurality of insurance sub-information based on the intermediate insurance sub-information corresponding to each recall unit, wherein the preset recall strategy of each recall unit is different.

[0085] Step S316: Calculate the correlation between each insurance sub-information in the plurality of insurance sub-information and the insurance analysis request.

[0086] Step S318: sort the multiple insurance sub-information according to the relevance to obtain a sub-information list.

[0087] Step S320: Input the sub-information list and the initial insurance information into the target reasoning model for processing to obtain the target insurance sub-information associated with the insurance analysis request.

[0088] In summary, in order to adapt to more insurance products and save human resources while improving the efficiency of insurance information analysis, the initial insurance information can be determined in response to the insurance analysis request submitted for the target insurance project; at this time, the initial insurance information can be input into the large language model, wherein the large language model includes an information update unit and a multi-way recall unit; in order to ensure the comprehensiveness of the analysis results, the initial insurance information can be updated to the target insurance information using the information update unit to avoid interference caused by redundant information, and then the multiple insurance sub-information associated with the insurance analysis request can be recalled in the target insurance information through the multi-way recall unit, thereby ensuring the comprehensiveness of the recall results, and then the multiple insurance sub-information and the initial insurance information can be input into the target reasoning model for processing, so that the reasoning model can be further analyzed from the semantic level to obtain the target insurance sub-information associated with the insurance analysis request. When analyzing insurance information, the large language model and the reasoning model can be combined to complete the information analysis, which can not only ensure the accuracy of the analysis, but also avoid the confusion caused by redundant information, thereby facilitating the use of downstream businesses.

[0089] Corresponding to the above method embodiment, this specification also provides an information processing device embodiment, Figure 4 FIG1 shows a schematic diagram of the structure of an information processing device provided by an embodiment of this specification. Figure 4 As shown, the device includes: A determination module 402 is configured to determine initial insurance information in response to an insurance analysis request submitted for a target insurance program; An input module 404 is configured to input the initial insurance information into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit; An updating module 406 is configured to update the initial insurance information to target insurance information using the information updating unit, and recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information using the multi-way recall unit; The processing module 408 is configured to input the multiple insurance sub-information and the initial insurance information into a target reasoning model for processing, and obtain target insurance sub-information associated with the insurance analysis request.

[0090] In an optional embodiment, determining the initial insurance information in response to an insurance analysis request submitted for a target insurance program includes: An insurance analysis request submitted for a target insurance project is received, and insurance information associated with the target insurance project is obtained according to the insurance analysis request; the insurance information is preprocessed, and the format of the preprocessed insurance information is converted to obtain initial insurance information.

[0091] In an optional embodiment, the updating the initial insurance information to the target insurance information by the information updating unit includes: The initial insurance information is input into the information updating unit; and the initial insurance information is semantically enhanced and rewritten using the information updating unit to obtain target insurance information.

[0092] In an optional embodiment, recalling the multiple insurance sub-information associated with the insurance analysis request from the target insurance information by the multi-channel recall unit includes: The target insurance information is retrieved through the recall strategy preset by each recall unit in the multiple recall units to obtain at least one initial insurance sub-information corresponding to each recall unit; the at least one initial insurance sub-information corresponding to each recall unit is fused to obtain the intermediate insurance sub-information corresponding to each recall unit; and multiple insurance sub-information are composed based on the intermediate insurance sub-information corresponding to each recall unit, wherein the recall strategy preset by each recall unit is different.

[0093] In an optional embodiment, inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request includes: Calculate the degree of association between each insurance sub-information in the multiple insurance sub-information and the insurance analysis request; sort the multiple insurance sub-information according to the degree of association to obtain a sub-information list; input the sub-information list and the initial insurance information into the target reasoning model for processing to obtain the target insurance sub-information associated with the insurance analysis request.

[0094] In an optional embodiment, after the step of inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain the target insurance sub-information associated with the insurance analysis request is performed, the method further includes: The initial insurance information and the target insurance sub-information are input into the evaluation model for processing to obtain an information evaluation result; if the information evaluation result is passed, the target insurance sub-information is fed back to the client that submitted the insurance analysis request; if the information evaluation result is failed, supplementary information is determined based on the information evaluation result, and the supplementary information and the target insurance sub-information are merged to obtain global insurance sub-information, and the global insurance sub-information is fed back to the client that submitted the insurance analysis request.

[0095] In an optional embodiment, after the step of inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain the target insurance sub-information associated with the insurance analysis request is performed, the method further includes: An adjustment request for the target insurance sub-information feedback is received, and model optimization information is constructed according to the adjustment request; and the large language model and the target reasoning model are optimized based on the model optimization information until a large language model and a target reasoning model that meet the optimization conditions are obtained.

[0096] The information processing device provided in this embodiment can determine initial insurance information in response to an insurance analysis request submitted for a target insurance project in order to adapt to more insurance products and save human resources while improving the efficiency of insurance information analysis. At this time, the initial insurance information can be input into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit. In order to ensure the comprehensiveness of the analysis results, the initial insurance information can be updated to the target insurance information using the information update unit to avoid interference caused by redundant information. Subsequently, the multiple insurance sub-information associated with the insurance analysis request can be recalled from the target insurance information through the multi-way recall unit, thereby ensuring the comprehensiveness of the recall results. Thereafter, the multiple insurance sub-information and the initial insurance information can be input into the target reasoning model for processing, so that the reasoning model can further analyze from a semantic level to obtain the target insurance sub-information associated with the insurance analysis request. When analyzing insurance information, the large language model and the reasoning model can be combined to complete the information analysis, which not only ensures the accuracy of the analysis but also avoids the confusion caused by redundant information, thereby facilitating downstream business use.

[0097] The above is a schematic diagram of an information processing device according to this embodiment. It should be noted that the technical solution of the information processing device and the technical solution of the above-mentioned information processing method are based on the same concept. For details not described in detail in the technical solution of the information processing device, please refer to the description of the technical solution of the above-mentioned information processing method.

[0098] Corresponding to the above method embodiment, this specification also provides another information processing device embodiment, Figure 5 FIG1 shows a schematic diagram of the structure of another information processing device provided by an embodiment of this specification. Figure 5 As shown, the device includes: A request receiving module 502 is configured to receive an insurance analysis request submitted by a user for a target insurance project, and obtain initial insurance information associated with the target insurance project according to the insurance analysis request; An input model module 504 is configured to input the initial insurance information into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit; an information updating module 506 configured to update the initial insurance information to target insurance information using the information updating unit, and recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information using the multi-way recall unit; An information processing module 508 is configured to input the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing, and obtain target insurance sub-information associated with the insurance analysis request; The information building module 510 is configured to build explanatory information corresponding to the target insurance sub-information and display the target insurance sub-information and the explanatory information to the user.

[0099] The information processing device provided in this embodiment can determine initial insurance information in response to an insurance analysis request submitted for a target insurance project in order to adapt to more insurance products and save human resources while improving the efficiency of insurance information analysis. At this time, the initial insurance information can be input into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit. In order to ensure the comprehensiveness of the analysis results, the initial insurance information can be updated to the target insurance information using the information update unit to avoid interference caused by redundant information. Subsequently, the multiple insurance sub-information associated with the insurance analysis request can be recalled from the target insurance information through the multi-way recall unit, thereby ensuring the comprehensiveness of the recall results. Thereafter, the multiple insurance sub-information and the initial insurance information can be input into the target reasoning model for processing, so that the reasoning model can further analyze from a semantic level to obtain the target insurance sub-information associated with the insurance analysis request. When analyzing insurance information, the large language model and the reasoning model can be combined to complete the information analysis, which not only ensures the accuracy of the analysis but also avoids the confusion caused by redundant information, thereby facilitating downstream business use.

[0100] The above is a schematic diagram of another information processing device according to this embodiment. It should be noted that the technical solution of the information processing device and the technical solution of the above-mentioned information processing method are based on the same concept. For details not described in detail in the technical solution of the information processing device, please refer to the description of the technical solution of the above-mentioned information processing method.

[0101] Figure 6 6 shows a block diagram of a computing device 600 according to one embodiment of the present disclosure. Components of the computing device 600 include, but are not limited to, a memory 610 and a processor 620. The processor 620 is connected to the memory 610 via a bus 630, and a database 650 is used to store data.

[0102] Computing device 600 also includes an access device 640 that enables computing device 600 to communicate via one or more networks 660. Examples of such networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. Access device 640 may include one or more of any type of network interface (e.g., a network interface card (NIC)) whether wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a Worldwide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, or a near field communication (NFC) interface.

[0103] In one embodiment of the present specification, the above components of the computing device 600 and Figure 6 Other components not shown in the figure may also be connected to each other, for example, via a bus. Figure 6 The computing device structure block diagram shown is for illustrative purposes only and is not intended to limit the scope of this specification. Those skilled in the art may add or replace other components as needed.

[0104] Computing device 600 can be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, personal digital assistant, laptop computer, notebook computer, netbook computer, etc.), a mobile phone (e.g., a smartphone), a wearable computing device (e.g., a smartwatch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or personal computer (PC). Computing device 600 can also be a mobile or stationary server.

[0105] The processor 620 is configured to execute the following computer-executable instructions, which implement the steps of the above-mentioned information processing method when executed by the processor.

[0106] The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the above-mentioned information processing method are based on the same concept. For details not described in detail in the technical scheme of the computing device, please refer to the description of the technical scheme of the above-mentioned information processing method.

[0107] An embodiment of the present specification further provides a computer-readable storage medium storing computer-executable instructions, which implement the steps of the above-mentioned information processing method when executed by a processor.

[0108] The above is a schematic diagram of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of the storage medium and the technical solution of the information processing method described above are based on the same concept. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solution of the information processing method described above.

[0109] An embodiment of the present specification further provides a computer program product, including a computer program or instructions, which implement the steps of the above-mentioned information processing method when executed by a processor.

[0110] The above is an illustrative solution of a computer program product of this embodiment. It should be noted that the technical solution of this computer program product and the technical solution of the above-mentioned information processing method are based on the same concept. For details not described in detail in the technical solution of the computer program product, please refer to the description of the technical solution of the above-mentioned information processing method.

[0111] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0112] The computer instructions include computer program code, which may be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium may include any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. It should be noted that the content of the computer-readable medium may be appropriately increased or decreased based on the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media does not include electric carrier signals and telecommunication signals.

[0113] It should be noted that for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of this specification are not limited by the order of the actions described, because according to the embodiments of this specification, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the embodiments of this specification.

[0114] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0115] The preferred embodiments disclosed above are intended only to help illustrate this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of the embodiments described herein. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the embodiments, thereby enabling those skilled in the art to better understand and utilize this specification.

Claims

1. An information processing method, characterized in that: include: determining initial insurance information in response to an insurance analysis request submitted for a target insurance program; Inputting the initial insurance information into a large language model, wherein the large language model includes an information updating unit and a multi-way recall unit; Using the information updating unit to update the initial insurance information to target insurance information, and using the multi-way recalling unit to recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information; The multiple insurance sub-information and the initial insurance information are input into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request.

2. The information processing method according to claim 1, wherein: The determining of initial insurance information in response to an insurance analysis request submitted for a target insurance program includes: receiving an insurance analysis request submitted for a target insurance project, and obtaining insurance information associated with the target insurance project according to the insurance analysis request; The insurance information is preprocessed and formatted to obtain initial insurance information.

3. The information processing method according to claim 1, wherein: The updating of the initial insurance information to target insurance information by the information updating unit includes: inputting the initial insurance information into the information updating unit; The information updating unit is used to perform semantic enhancement and information rewriting on the initial insurance information to obtain target insurance information.

4. The information processing method according to claim 1, wherein: The recalling of the plurality of insurance sub-information associated with the insurance analysis request from the target insurance information by the multi-channel recall unit includes: Performing information retrieval on the target insurance information using a recall strategy preset by each of the multiple recall units to obtain at least one initial insurance sub-information corresponding to each recall unit; Fusing at least one initial insurance sub-information corresponding to each recall unit to obtain intermediate insurance sub-information corresponding to each recall unit; A plurality of insurance sub-information is formed based on the intermediate insurance sub-information corresponding to each recall unit, wherein the recall strategy preset for each recall unit is different.

5. The information processing method according to claim 1, wherein: Inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request includes: calculating a correlation degree between each insurance sub-information in the plurality of insurance sub-information and the insurance analysis request; Sort the multiple insurance sub-information according to the relevance to obtain a sub-information list; The sub-information list and the initial insurance information are input into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request.

6. The information processing method according to any one of claims 1 to 5, characterized in that: After the step of inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request is performed, the method further includes: Inputting the initial insurance information and the target insurance sub-information into an evaluation model for processing to obtain an information evaluation result; If the information evaluation result is passed, feeding back the target insurance sub-information to the client that submitted the insurance analysis request; If the information evaluation result is failed, supplementary information is determined based on the information evaluation result, and the supplementary information is merged with the target insurance sub-information to obtain global insurance sub-information, which is then fed back to the client that submitted the insurance analysis request.

7. The information processing method according to any one of claims 1 to 5, characterized in that: After the step of inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request is performed, the method further includes: receiving an adjustment request for the target insurance sub-information feedback, and constructing model optimization information according to the adjustment request; The large language model and the target reasoning model are optimized based on the model optimization information until a large language model and a target reasoning model that meet optimization conditions are obtained.

8. An information processing method, characterized in that: include: Receiving an insurance analysis request submitted by a user for a target insurance project, and obtaining initial insurance information associated with the target insurance project according to the insurance analysis request; Inputting the initial insurance information into a large language model, wherein the large language model includes an information updating unit and a multi-way recall unit; Using the information updating unit to update the initial insurance information to target insurance information, and using the multi-way recalling unit to recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information; Inputting the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing to obtain target insurance sub-information associated with the insurance analysis request; Construct explanatory information corresponding to the target insurance sub-information, and display the target insurance sub-information and the explanatory information to the user.

9. An information processing device, characterized in that include: a determination module configured to determine initial insurance information in response to an insurance analysis request submitted for a target insurance program; an input module configured to input the initial insurance information into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit; an updating module configured to update the initial insurance information to target insurance information using the information updating unit, and recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information using the multi-way recall unit; The processing module is configured to input the multiple insurance sub-information and the initial insurance information into a target reasoning model for processing, and obtain target insurance sub-information associated with the insurance analysis request.

10. An information processing device, characterized in that: include: a request receiving module configured to receive an insurance analysis request submitted by a user for a target insurance project, and obtain initial insurance information associated with the target insurance project according to the insurance analysis request; An input model module is configured to input the initial insurance information into a large language model, wherein the large language model includes an information update unit and a multi-way recall unit; an information updating module configured to update the initial insurance information to target insurance information using the information updating unit, and recall multiple insurance sub-information associated with the insurance analysis request from the target insurance information using the multi-way recall unit; an information processing module configured to input the plurality of insurance sub-information and the initial insurance information into a target reasoning model for processing, and obtain target insurance sub-information associated with the insurance analysis request; The information construction module is configured to construct explanation information corresponding to the target insurance sub-information and display the target insurance sub-information and the explanation information to the user.

11. A computing device, characterized in that include: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer-readable storage medium, characterized in that It stores computer-executable instructions, which, when executed by a processor, implement the steps of the method according to any one of claims 1 to 8.

13. A computer program product, characterized in that The method comprises a computer program or an instruction, which implements the steps of the method according to any one of claims 1 to 8 when the computer program or the instruction is executed by a processor.

Images